Analyse phénotypique des lignées transgéniques sélectionnées

Publication n° 3

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

Two EguCBF1 genes overexpressed in Eucalyptus

display a different impact on stress tolerance and plant development

Navarro, M.1, Ayax C.1, Martinez Y.1, El Kayal, W.2, Marque C.1 and Teulières C1.

1 : Université de Toulouse (UT3) : ERT 1045, Pôle de Biotechnologie Végétale, 24 Chemin de Borde Rouge BP 42617 Auzeville, 31326 Castanet-Tolosan, France.

2 : University of Alberta, Edmonton, Alberta, Canada

All correspondence should be addressed to: Chantal TEULIERES:

Phone: 0033 562 193 522 Fax: 0033 562 193 502

e-mail: teulieres@scsv.ups-tlse.fr

Total word count: 6430 Introduction: 639

Materials and Methods: 1357 Results: 2263

Discussion: 2106 Acknowledgements: 65

Number of figures: six Number of table: four

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

Summary

Two C-repeat Binding Factor genes (EguCBF1a/b), isolated from E. gunnii and differentially cold-regulated were constitutively overexpressed in a cold sensitive Eucalyptus hybrid. The resulting transgenic lines (EguCBF1a-OE and EguCBF1b-OE) exhibited a very altered phenotype including improvement of constitutive cold tolerance and cold acclimation capacity, a decrease in stomata density and an over-accumulation of anthocyanins. These characteristics were also observed to a lesser extent in a cold-acclimated control plant, confirming that, for Eucalyptus, CBF overexpression mimics cold acclimation. Moreover, the induction of five putative CBF target genes was observed in CBF-overexpressing lines as well as in the cold-acclimated control line.

Compared to the control plant, the most altered transgenic line (EguCBF1a-OE A1 line), exhibited better dehydration tolerance, higher oil gland density and a wax deposition on leaves. Its growth, cell volume, and the area and thickness of the leaves were also significantly reduced. Surprisingly, the EguCBF1b-OE B9 line, with a level of transgene expression equivalent to the A1 line, showed a less marked phenotype, suggesting a difference in stability and/or transactivation efficiency between EguCBF1A and B factors. The features of these transgenic lines, provide the first signs of adaptive mechanisms controlled by CBF transcription factors in an evergreen broad-leaved tree.

Key words : CBF overexpression, transgenic Eucalyptus, stress tolerance, reduced growth, stomata, oil gland, anthocyanin, wax

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

Introduction

Cold affects both the productivity and quality of crops and severely limits their growth range. Plant survival in frost conditions is very closely associated with their ability to adjust to temperature changes and trigger rapid protective responses that result in a transient increase in freezing tolerance. This phenomenon, called cold acclimation, occurs after exposure to non- freezing temperatures and involves a number of alterations at the molecular and cellular level. Most of these changes, which are mainly directed toward membrane and protein protection, are controlled by transcriptome modifications. The cold-induced genes encode functional proteins such as dehydrins, chaperones or the enzymes involved in the synthesis of osmoprotectants. A number of these genes which contain the CRT/DRE cis-acting element are involved in the response to dehydration, high-salt and low-temperature through induction of DREB genes including DREB1/CBF and DREB2 (Liu et al., 1998). These effector genes, upregulated in transgenic lines overexpressing CBF genes, belong to the CBF regulon (Jaglo- Ottosen et al., 1998).

CBF genes have been identified in all the 80 higher plant species (cold tolerant and sensitive) studied to date including 25 woody plants. There is now compelling evidence that the CBF pathway plays a prominent role in freezing tolerance since the overexpression of a CBF gene improves this trait. Natural variability in cold tolerance also appears to be partly mediated through the CBF pathway, as suggested in wheat, Arabidopsis or Citrus (Hannah et al., 2006; Miller et al., 2006; Champ et al., 2007). In cold sensitive species, CBF induction might be later, shorter or weaker. Alternatively, the CBF proteins might be less efficient or the CBF regulon size smaller (Zhang et al., 2004; Xiao et al., 2008).

Eucalyptus is an evergreen broad-leaved tree without endodormancy that is characterized by opportunistic growth mainly controlled by mean temperature and available water. Highly exposed to winter frosts over long periods, this tree may be considered as an interesting system for studying cold tolerance mechanisms without interference from the endodormancy process. E. gunnii, which is able to withstand frosts of up to -18°C, is one of the most tolerant species of Eucalyptus. Four CBF paralogs were previously isolated from E. gunnii and RT-qPCR analysis highlighted complementary expression profiles in a range of natural standard and cold conditions (El Kayal et al., 2006; Navarro et al., 2009). Besides this regulation specificity, these genes may also differ in their protein function as suggested in several plants (Gao et al., 2002; Novillo et al., 2007; Xiao et al., 2008).

To explore the function of the different EguCBF1 genes, a functional study was undertaken. To date, a number of CBF-overexpressing transgenic plants have been described

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

(Agarwal et al., 2006). As well as the impact on the tolerance to cold and other osmotic stresses, CBF overexpression often resulted in alterations in growth and development, the mechanisms of which have not yet been completely elucidated. However, most of these studies deal with heterologous transformations in model systems, leaving aside the identification of the corresponding CBF regulon and the investigation of the specific impact of the transformation on the phenotype. No overexpression of CBF genes from trees has been reported in homologous systems to date, probably due to the lack of efficient transformation procedures. Among the CBF gene families isolated from woody plants, only DREB1 from Cerasus has been overexpressed in Arabidopsis (Kitashiba et al., 2004).

In this paper, we present the generation and the phenotypic analysis of transgenic Eucalyptus lines overexpressing the EguCBF1a or EguCBF1b gene from E. gunnii. The transformation was performed on E. urophylla x E. grandis, a hybrid which is grown intensively in Brazil and which is more cold-sensitive than E. gunnii. The first objective was to evaluate the effect of CBF overexpression on cold tolerance and development of this tree. This study also explored whether the EguCBF1A and B proteins are equivalent for mediating cold response.

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

Materials and methods

Plant material

E. urophylla x E. grandis (clone 201) and the generated transgenic lines were micropropagated on MS medium supplemented with vitamins and hormones (M medium) as previously described (Tournier et al., 2003). The microcuttings were grown in a controlled- environment chamber at 25°C-day/22°C-night, under a long-day photoperiod (16h/light=115µmol.m-2.s-1 supplied by Lumilux Daylight 58W Osram). Every three weeks they were transferred onto a fresh medium.

Construction of transformation vectors

The transformation vectors were constructed using the gateway technology, according to the manufacturer’s instructions (Invitrogen). PCR fragments corresponding to the coding region of the EguCBF1a and EguCBF1b genes (665 bp and 674 bp respectively) were cloned into the plasmid pDNOR 201 by the standard BP clonase II reaction (Invitrogen) (El Kayal et al., 2006). Using LR clonase (Invitrogen), the entry clones and the destination vector pK7WG2D were associated to construct each expression vector (Karimi et al., 2002). The final constructs (pK7, pCBF1a, pCBF1b) were introduced by heat shock into Agrobacterium tumefaciens (strain AGL1) before the plant transformation experiments.

The three resulting transformation constructs contained the nptII gene, which confers kanamycin resistance and the gfp gene, which is a reporter gene used for the selection of transformants. The transformation vectors that housed a CBF coding sequence (EguCBF1a or b) controlled by the 35SCaMV promoter were used to generate transgenic lines overexpressing a CBF gene. The control transgenic lines were generated using pK7 vector lacking the CBF gene.

Generation and screening of transgenic lines

The genetic transformation of E. urophylla x E. grandis leaves, isolated from microcuttings, was achieved via A. tumefaciens (AGL1) as previously described (Tournier et al., 2003). The selection of transgenic lines based on their kanamycin resistance was confirmed by the observation of fluorescence due to the gfp expression in transgenic sectors, which were isolated from explants before subculture.

Finally, the presence of transferred sequence (T-DNA) in genomic DNA of each GFP- positive lines was demonstrated through PCR experiments conducted using specific primers (nptII = GCGGTTCTGTCAGTTCCAAACGTAA;

Table 1 : List of oligonucleotide sequences used in RTq-PCR experiments

Gene Name Oligonucleotide sequences ADNc matrix

Fw 5’-CCCTTTCTCTTCTCATTCCCAT-3’ EguCBF1a Rev 5’-AGAAGTTCCCAAGTGCCGTG-3’ 240 ng Fw 5’-CTCTTCCTCTTATATCTCCCATCCC-3’ EguCBF1b Rev 5’-TCCTGTCGTGGGACAGCAG-3’ 240 ng Fw 5’-CGCGCTACACTGATGTATTC-3’ 18S Rev 5’-GTACAAAGGGCAGGGACGTA-3’ 0.24 ng Fw 5’-TGCCAAAATAAGTGCAAGGTC-3’ DHN1 Rev 5’-TGCGAAGGACTCAGTACACAA-3’ 24 ng Fw 5’-AGTGCAAGAAGGCCGCTACTT-3’ DHN2 Rev 5’-GAACAACTTTCAGGTGCAAGCTT-3’ 240 ng Fw 5’-GCAGGGTGTCTGGATCTGGTT-3’ Thioredoxin Rev 5’-CAAAAGCTCTCACTGGAACAACA-3’ 240 ng Fw 5’-TTGAGTGGCAACACATGATGG-3’ Lti6 Rev 5’-ATCCCACGAAATCGCGACT-3’ 24 ng Fw 5’-TGCATGTGCTGTGAAAGTCA-3’ Metallothionein Rev 5’-CCACGGAAGAAGCACAAAAG-3’ 24 ng Fw 5’-GATTGAGGAGGTCGACTAAGCG-3’ HSP70 Rev 5’-CGAAATGCAAAAGACCTGGC-3’ 24 ng

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

gfp = AATGATAGGAGAAGTGAAAAGATGA).

Gene transcript measurements through quantitative RT-qPCR

The aim of the transcript quantification was to specifically measure the expression of the EguCBF1a or b transgene and six putative EguCBF1 target genes which had previously been isolated from Eucalyptus (Keller et al., 2009).

Total RNA was extracted from E. urophylla x E. grandis transgenic lines (three extracts per line were obtained from microcuttings) using the SV Total RNA Isolation System (Promega, France). Using SuperScript II and random primers (Invitrogen, France), cDNAs were produced according to the manufacturer’s instructions.

To quantify the transcript level of transgenes, specific primers corresponding to EguCBF1a/b coding sequences were designed (Tab. 1) using the Primer Express software (version 2.0, Applied Biosystems, France). The PCRs were performed in 10µl of 2X SYBR Green power plus Master Mix (Applied Biosystems), using 300nM of each primer and 240ng of cDNA for the EguCBF1a/b genes. Three replicates of each PCR were run through an ABI PRISM 7900HT Sequence Detection System (Applied Biosciences, France) as previously described (El Kayal et al., 2006). Specific primers for 18S RNA were used as an internal control for the normalization of the RNA steady-state level, and the relative changes in gene expression were quantified using the 2-∆∆Ct method (Livak and Schmittgen, 2001). The results of the EguCBF1a/b transgene relative transcript abundance were presented as a mean value of the three assay replicates of the transgenic line compared with the mean of the three control values (leaves from control plants), corresponding to endogenous expression of each EguCBF1.

The same procedure was used to measure relative transcript abundance of EguCBF1 putative target genes on 240ng cDNA for the Thioredoxin or DHN2 genes and on 24ng for DHN1, Lti6b, Metallothionein or HSP70 genes.

Freezing and dehydration tolerance assessment

The transgenic lines overexpressing EguCBF1a or b genes were compared to the control transgenic line (pK7) for freezing tolerance before and after a cold acclimation program corresponding to 18 days of culture at chilling temperatures.

The results were obtained from three independent experiments using at least four microcuttings from each transgenic line. They were transferred to Falcon tubes (50ml) containing 15ml of M medium and grown in a controlled-environment chamber at 25°C-

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

day/22°C-night, with a long-day photoperiod (16h/light=115µmol.m-2.s-1 supplied by Lumilux Daylight 58W Osram) for one week. For cold acclimation, the chilling program consisted of three steps: four days at 12°C/day and 8°C/night, then seven days at 8°C/day and 4°C/night and lastly, seven days at 4°C day and night. This program was coupled with a short photoperiod (8 h/light).

To evaluate frost tolerance, the microcuttings were exposed to a progressive, controlled freezing program, consisting of two steps: firstly, the temperature decreased from 22°C to -1°C over 24h, and then was held at -1°C for 16h. Secondly, the microcuttings were exposed to freezing temperatures from -1°C down to -7.5°C, -8°C or -8.5°C at a speed of 1°C/h. Then, the treated microcuttings were cultured in standard conditions overnight before being transferred into fresh M medium. Finally, after five days under control conditions, the lines were scored for freezing tolerance, based on the recovery capacity of the microcuttings, and expressed as the survival rate (%).

Dehydration was induced after removing explants from the culture medium, and placing them on Whatman paper at 23°C in the dark. Five microcuttings per line were weighed every hour during eight hours. The resulting values for each line were converted to the percentage of their corresponding initial fresh weight. The loss of fresh weight probably corresponds to the water-loss representing the dehydration tolerance.

Measurement of microcutting growth and leaf size

The production of biomass was evaluated after three weeks of culture corresponding to a micropropagation cycle. Seven microcuttings from each tested line (pK7, A1 and B9 CBF- OE lines) were first weighed before being transferred into fresh medium and then after three weeks of growth. The results are presented as a percentage of initial weight for each line.

All the leaves from two microcuttings per line were detached and collected on a paper before scanning. The images were then analysed using the Image Pro Plus (4.5.0.29) software to determine the length and width of each leaf and its resulting area. After measuring between 90-150 leaves per line, the corresponding leaf areas were determined and were then classified into three groups: small (0-2 mm2), medium (2-8.5 mm2) and large (8.5-25 mm2). The results are expressed for each line as the distribution (%) of these three classes of areas.

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

Microscopic analysis of leaves

Young or old leaf samples collected from two microcuttings were fixed in 2.5% glutaraldehyde in 0.05 M sodium cacodylate buffer, pH 7.2, for 24 h at room temperature and then dehydrated in an ethanol series (20, 40, 60, 80, and 100%). For Scanning Electron Microscopy (SEM), they were subjected to critical point drying (CPD 750 EmScope), coated with gold-palladium, and observed with a Hitachi S450 scanning electron microscope (Hitachi, Ibraki, Japan).

For the histological analysis, four leaves were collected from each of two microcuttings of CBF1-OE and control lines. The leaves were embedded in LR White and 1µm sections (24 cross-sectional segments per leaf) were mounted on glass slides. The sections were stained with 0.01M toluidine blue dissolved in citrate buffer (pH 4.2) for 1min 30s. Leaf thickness was visualized with a Leica DMIRBE microscope and measured using Image pro plus (4.5.0.29) software.

On randomly chosen microscopic views corresponding to 1mm2, the mesophyll pavement cells, stomata and oil gland scores were used to deduce the corresponding densities: Ed, Sd and OGd respectively. The index value for stomata (Si), corresponding to the percentage of stomatal cells, was calculated for the control and OE lines using the following formula : Si [%] = [(Sd)/(Sd + Ed)] x 100 (McCauley and Evert, 1988); the same calculation was used to obtain the index of oil glands.

Anthocyanin content measurement

At least three microcuttings (0.2g x3) carried out on triplicate samples were ground in liquid nitrogen. Anthocyanins were extracted as previously described (Close et al., 2000) before absorbance measurements at 530 and 657nm. The corrected value of absorbance was calculated (A530-0.25A657) to remove the absorbance of chlorophyll and degradation products

(Mancinelli et al., 1975). The results are expressed as anthocyanin content per gram of fresh weight.

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

Results

Molecular screening of transgenic lines

From the detached leaves of E. urophylla x E. grandis (clone 201) incubated with A. tumefaciens, independent lines were regenerated on selective medium containing kanamycin. During this regeneration step, 30 potentially transformed sectors (calli or buds) were isolated after observation of GFP fluorescence areas on explant surfaces. The presence of T-DNA sequence in the genome of these GFP-positive lines was confirmed through PCR experiments (data not shown) using primers specifically designed to amplify the complete T-DNA sequence, from nptII to gfp genes. Among them, 16 lines were obtained with pCBF1a vector, 11 with pCBF1b vector and three with pK7 (control vector).

The final step of the molecular screening of the transgenic lines was achieved using RT-qPCR with specific primers designed on the coding region of the EguCBF1a or b gene. Considering the EguCBF1a/b endogenous transcript level of the transgenic control line (pK7) as the control value (=1), this experiment led to the identification of 18 CBF-overexpressing lines. They exhibited a transgene level ranging from 1390 to 6002 (7 CBF1a-OE lines) and from 2520 to 5560 (11 CBF1b-OE lines) respectively.

Freezing tolerance of transgenic lines before or after cold acclimation

The constitutive freezing tolerance level of the control transgenic line (pK7) at -7.5°C was found to be similar (~30%) to the non-transformed line. The first evidence of the positive impact of CBF1 overexpression on constitutive freezing tolerance was provided by the CBF1- OE lines which survived at -8°C while the control lines died.

Three CBF1a-OE lines exhibiting a high transgene level (3165, 4309 and 6002) and the best freezing tolerance (75% survival rate at -8°C) could not be selected for further analyses because they also presented a severe growth alteration which was illustrated by the microcuttings’ reduced capacity to multiply. Apart from these three lines, the phenotypic modifications observed for the CBF-OE lines did not compromise the microcutting development.

Overall, the freezing tolerance, expressed as the survival rate at -8°C, was found to improve for the 18 transgenic lines overexpressing either the CBF1a or b gene (data not shown). From the CBF1a-OE lines selected on transgene expression level (from 1390 to 6002), the freezing tolerance strongly increased for five lines (63-75% survival rate at -8°C) and moderately increased for two lines (25-42%). On the other hand, for the 11 CBF1b-OE

Figure 1: Freezing tolerance and CBF transgene transcript abundance of the transgenic lines.

The freezing tolerance at -8°C of the transgenic lines overexpressing EguCBF1a or b genes was compared to the control line (pK7). The presented data were obtained from three experiments using at least four microcuttings per line exposed during three days to a progressive freezing program. The recovery capacity of these microcuttings after five days in a fresh culture medium was evaluated and presented as the survival rate (dark histograms). Abundance of EguCBF1a/b transcripts was quantified in comparison to 18S RNA transcript level, used as an internal control for normalization of RNA steady-state level. The results of EguCBF1a/b transgene relative abundance correspond to the mean value of three assay replicates for each CBF-OE line compared to the mean of the three values for endogenous EguCBF1a/b transcript quantity in the control line. A representative histogram with standard deviation from three replicates performed for each point has been represented (white histograms).

Table 2: Freezing tolerance of the CBF-OE and control lines after cold acclimation. Survival rate (%) after freezing

-7.5°C -8°C -8.5°C pK7 30.8 0 0 pK7 acclimated nd 66.6 0 A1 100 66.7 0 A1 acclimated nd nd 77.7 A25 100 41.7 0 A25 acclimated nd nd 50 A21 50 25 0 A21 acclimated nd nd 50 B9 87.5 50 0 B9 acclimated nd nd 22.2 B8 75 33.3 0 B8 acclimated nd nd 16.6 B14 55.6 12.5 0 B14 acclimated nd nd 11.1

The microcuttings were cultured following the 18-day chilling program (see Material and Methods for details) and then exposed to freezing temperatures (-7.5°C, -8°C or -8.5°C). The recovery of microcuttings was expressed in term of survival rate, as described for constitutive freezing tolerance (Fig. 1).

Chapitre 2 - Surexpression des gènes EguCBF1a ou 1b chez E. urophylla x E. grandis et Arabidopsis thaliana

lines presenting a high transgene expression level (from 2520 to 5560), a lower impact on freezing tolerance was observed: five lines were significantly more tolerant than the control (25-50% survival rate at -8°C) compared to the five other lines which were only slightly improved (<17%), while the last one was similar to the control line.

A representative set of six lines that exhibited a high, medium or low level of transgene expression was finally chosen. As shown in Fig. 1, all of them display increased freezing tolerance in comparison with the transgenic control line (pK7). The three selected CBF1a-OE lines (A1, A25 and A21) showed a transgene expression level that was correlated quite well with freezing tolerance. The situation was different when CBF1b-OE selected lines (B8, B9 and B14) were considered: for a similar transgene expression level (5114 and 4816), the B8 and B9 lines exhibited contrasted survival rates at -8°C (respectively 33.3 and 50%). Interestingly, the abundance of transgene transcripts was similar for the B9 and A1 lines (4816 and 4724 respectively) which harbour the two different EguCBF1 genes under study. However, the A1 line exhibited a higher survival rate after freezing (66.7%) than the B9 line (50%).

The improvement in constitutive freezing tolerance due to overexpression is also demonstrated by the change of the minimum tolerated temperature, from -7.5°C to -8°C in standard culture conditions (Tab. 2). To investigate the cold acclimation capacity of the OE and control lines, the microcuttings were submitted to the progressive chilling program (for details, see Material and Methods). The increase in freezing tolerance after this cold culture